JP2006195644A - Device and method for executing virtual machine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To speed up arithmetic processing in execution of a virtual machine, while simplifying the configuration. <P>SOLUTION: In the virtual machine execution device adapted to load an intermediate cord from a secondary storage to a main storage followed by execution, a virtual machine engine includes a class sequence including a class the intermediate code of which is loaded to the main storage, and a run-time sequence including a runtime storing information in execution. Each class includes a method member having a method describing processing procedure information including an operation code and a field member. When the intermediate code is loaded to the main storage, the address of the operation code contained in the method is expressed by the address of the main storage and stored in the data area of each class. The runtime includes a local variable area storing variables and a stack area temporarily storing a calculation result in processing including an operand, and the operand stored in the stack area is expressed by the corresponding data format and value form. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a virtual machine execution apparatus and method for executing an intermediate code independent of a hardware architecture of a computer and executing a virtual machine such as a Java virtual machine.

  The intermediate code method converts from a surface language that can be understood by humans to an executable code that can be understood by a machine, and then converts it into an intermediate code that is a predetermined code system. This is a method that implements processing while reading and interpreting code. These technologies have become commonplace with the advent of Java technology (where Java is a registered trademark of Sun Microsystems). The intermediate code execution device is sometimes called a “virtual machine”, and the intermediate code is sometimes called a “byte code”.

  Non-Patent Document 1 describes the specifications of a Java virtual machine, and Chapter 6 of Non-Patent Document 1 defines a bytecode instruction set. For example, regarding arithmetic instructions, instructions such as (iadd / ladd / fadd / dadd) are provided for addition, and (imul / lmul / fmul / dmul) are provided for multiplication. As described above, since instructions for integers and floating-point numbers have different behaviors in overflow and division by zero, the conventional virtual machine prepares instructions for each.

  Further, the same program can be distributed to different types of computers depending on the intermediate code system. While some computers such as workstations and personal computers have powerful processing capability and computers with sufficient memory, they are also used in CPUs with inferior processing capability and computers with limited memory, such as embedded systems. Further, in many cases, an intermediate code execution device such as a virtual machine is realized by an interpreter method, and there is a problem that an execution speed is slower than an execution speed by a native code.

  In order to solve this problem, Japanese Patent Application Laid-Open No. 2005-228561 discloses that “in order to provide a compiler capable of executing Java bytecode at high speed with limited resources, the bytecode is an address of a processing routine corresponding to each bytecode. Is converted to an intermediate code having an operation code, and the processing routine is executed in the order of the intermediate code, and at this time, the intermediate code is set to a fixed length, and the bit position of the operation code is also fixed therein. The offset from the start address of the processing routine group 104 is used for the operation code, and the next intermediate is used in parallel with the original processing for the intermediate code by utilizing the CPU pipeline free space. Part of the instruction group for jumping to the code processing routine (operand generation and jump), and the next intermediate code A part of the instruction group for jumping to the processing routine of the host (execution of prefetch of intermediate code and address calculation of the processing routine) is executed. ”That is, the device disclosed in Patent Document 1 (hereinafter referred to as the first conventional example) )), The intermediate code string is converted to an internal intermediate code string that can be executed at the time of loading at a high speed. Specifically, the intermediate code conversion process at the time of loading corresponds to each instruction. An offset from the reference address of the instruction execution module is calculated and converted. At the time of execution, the obtained offset and the reference address are added to calculate the instruction execution module start address.

  Further, Patent Document 2 states that “in an intermediate code execution device that executes intermediate code that does not depend on the hardware architecture of a computer, the execution of the intermediate code is accelerated while the register of the central processing unit is used for the purpose of speeding up. In order to avoid the increase in the storage area to be used and to prevent an increase in the used storage area, each instruction in the input intermediate code string is internally replaced by a replacement means that replaces the instruction execution module end address with the relative address from the instruction execution module end address. An intermediate code string is prepared, and the internal intermediate code is sequentially executed by the relative address acquisition unit and the relative jump unit using the relative address. That is, the intermediate code execution device disclosed in Patent Document 2 (hereinafter referred to as the second conventional example) has a table of the start address and end address of each module, and shifts from the current module to the next module. Then, a relative address is calculated from the end address of the current module and the start address of the next module, and a relative jump is performed to move to the next module.

  Further, in Patent Document 3, “in order to provide an information processing apparatus capable of executing calculations for different data types based on a common algorithm with the same execution program, the data type of data to be operated is indicated. A register and a means for setting a data type in the register, a common instruction means for at least one of the plurality of data types, and corresponding to the contents of the instruction means and the register By performing data type operations, calculations for different data types based on a common algorithm are executed by the same execution program ”. That is, in the information processing apparatus disclosed in Patent Document 3 (hereinafter, referred to as a third conventional example), by preparing a register in a compiler apparatus that converts a high-level language source program into a machine language program and outputs it, The compiler device itself is simplified.

Japanese Patent Laid-Open No. 11-175349. JP2003-216443. Japanese Patent Laid-Open No. 60-169944. Published by Tim Lindholm et al., Translated by Masaaki Murakami, “Java Virtual Machine Specification 2nd Edition”, Pearson Education, May 2001.

  Since the behavior differs between an integer and a floating-point number, an intermediate code execution device according to the prior art including a Java virtual machine prepares an instruction for each. An instruction set corresponding to the data type is required for all arithmetic instructions such as addition, subtraction, multiplication, division, remainder, and sign inversion, so the table size increases and the intermediate code execution device is made compact. It was difficult. In the technique disclosed in the first conventional example, it is necessary to calculate the instruction execution module start address by adding the reference address and the acquired offset at the time of execution. Therefore, there has been a problem that it still requires arithmetic processing at the time of execution.

  In addition, in the intermediate code execution device according to the second conventional example, a technique for increasing the execution speed is disclosed, but as described above, a table of the start address and end address of each module is provided. When moving from the current module to the next module, the relative speed is calculated from the end address of the current module and the start address of the next module, and the relative speed is used to move to the next module. There was a problem that could not be improved significantly. Furthermore, in the information processing apparatus according to the third conventional example, the configuration of the compiler can be simplified, but the configuration of the interpreter type virtual machine cannot be simplified.

  An object of the present invention is to provide a virtual machine execution apparatus and method that can solve the above problems and can speed up the arithmetic processing at the time of execution of the virtual machine as compared with the prior art.

  Another object of the present invention is to provide a virtual machine execution apparatus and method capable of solving the above-described problems and simplifying the configuration of the virtual machine as compared with the prior art.

A virtual machine execution device according to a first aspect of the present invention is a virtual machine execution device that loads and executes an intermediate code from a secondary storage device to a main storage device.
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
When the intermediate code is loaded into the main storage device, the address of the operation code included in the method is expressed by the address of the main storage device and stored in the data area of each class.

In the virtual machine execution device, the runtime can secure a local variable area for storing an address to an execution method and a variable generated and used only in the method, and temporarily calculates a calculation result during the process. With a stack area to store
Operands expressed in a data type and a value format are stored in the stack area.

A virtual machine execution device according to a second aspect of the present invention is a virtual machine execution device that loads an intermediate code from a secondary storage device to a main storage device and executes the intermediate code.
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
Operands expressed in a data type and a value format are stored in the stack area.

According to a third aspect of the present invention, there is provided a virtual machine execution method in which an intermediate code is loaded from a secondary storage device to a main storage device and executed.
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
When the intermediate code is loaded into the main memory, the address of the operation code included in the method is represented by the address of the main memory and stored in the data area of each class.

In the virtual machine execution method, the runtime can secure a local variable area for storing an address to an execution method and a variable generated and used only in the method, and temporarily calculates a calculation result during the process. With a stack area to store
Operands expressed in a data type and a value format are stored in the stack area.

According to a fourth aspect of the present invention, there is provided a virtual machine execution method in which an intermediate code is loaded from a secondary storage device to a main storage device and executed.
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
Operands expressed in a data type and a value format are stored in the stack area.

  Therefore, according to the present invention, when the intermediate code is loaded into the main storage device, the address of the operation code included in the method is represented by the address of the main storage device and stored in the data area of each class. Since it is not necessary to perform address translation processing when the virtual machine is executed, the virtual machine can be executed at a higher speed than in the prior art.

  Further, according to the present invention, since the operand stored in the stack area is expressed in the form of its data type and its value, a virtual machine having a compact specification can be suppressed without increasing the instruction code due to the combination of the data types. Can be realized.

  Hereinafter, embodiments according to the present invention will be described with reference to the drawings. In addition, the same code | symbol is attached | subjected about the same component.

Embodiment.
FIG. 1 is a block diagram showing a configuration of a virtual machine execution apparatus 1000 according to an embodiment of the present invention. A virtual machine execution device 1000 according to the present embodiment is a virtual machine execution device that executes intermediate code that does not depend on the hardware architecture of a digital computer. As shown in FIG. 1, as shown in FIG. The storage device 1200 includes a main storage device 1300, and these devices 1100, 1200, and 1300 are connected via a bus, and the secondary storage device 1200 has a class intermediate code 1201 that describes the execution contents in advance. Stored.

  FIG. 2 is a flowchart showing a virtual machine installation process executed by the central processing unit 1100 of FIG. 4 is a block diagram showing the configuration of the virtual machine execution device 1000 after the process of step S1 in FIG. 2, and FIG. 5 is a block diagram showing the configuration of the virtual machine execution device 1000 after the process of step S3 in FIG. FIG.

  2, first, in step S1, as shown in FIG. 4, the virtual machine engine 1400 and the processing routine group 1410 included in the class intermediate code 1201 are loaded from the secondary storage device 1200 to the main storage device 1300. Here, the loaded virtual machine 1400 has a class array for registering classes to be loaded later. A class is software that has at least one member variable and at least one member function, and is a collection of variables and functions that are somewhat conceptually related. The processing routine group 1410 is a basic process provided by the virtual machine 1400, and a number is assigned to each process. Specifically, the member function in the class is an array in which the numbers assigned in the processing routine group 1410 to be executed and the parameters corresponding thereto are arranged in the order of execution. In the example of FIG. 4, the processing routine group 1410 includes an operation code a execution module 1410a, an operation code b execution module 1410b, an operation code c execution module 1410c, and the like.

  Next, in step S2, as shown in FIG. 5, the class file group included in the class intermediate code 1201 and used when the virtual machine is executed is loaded from the secondary storage device 1200 into the main storage device 1300. The secondary storage device 1200 stores at least one class. Instead of loading all classes, load only the classes you need. The user or other system that uses the virtual machine first specifies which member function of which class is the activation function. As a result, the class to be loaded from the secondary storage device 1200 is determined and loaded. A class may use functions of other classes, and the information is described in a file to be loaded. Based on this information, further required classes are loaded.

  In step S3, the class file group loaded in the main storage device 1300 performs link processing by replacing the locations that are referred to by name with the actual memory addresses on the main storage device 1300. Ends execution preparation processing by the virtual machine. In step S3, the link operation is executed when loading of all necessary classes is completed. In the state stored in the secondary storage device 1200, the other class reference information in the class is described by character string information such as a class name and a function name. When loading from the secondary storage device 1200 to the main storage device 1400 is completed, the address of the class in the main storage device 1400 becomes clear. Therefore, processing for replacing with the address in the main storage device 1400 is executed based on the other class reference information. .

  That is, in the main storage device 1300, as shown in FIGS. 4 and 5, there is a virtual machine engine 1400 that manages all information when the virtual machine is executed, and the virtual machine engine 1400 has a class array and a runtime array. The virtual machine engine 1400 loads the class intermediate code 1201 that is binary data of the intermediate code, from the secondary storage device 1200 to the main storage device 1300 as much as the code necessary for execution, and generates at least one class 1500 as an array. to manage. The runtime array is generated to store necessary information at the stage of actually executing processing from a predetermined activation method. Here, the class 1500 is a program module in which data and functions based on a certain concept are grouped into one group, and the class 1500 describes processing procedure information including an operation code as shown in FIG. A method member 1510 having at least one method, and a field member 1520 having at least one field data (including at least one operand) including at least one reference information to numeric data or character string data or other classes; As will be described later, a data area including the address of the operation code is included. A class array is an array having one or more such classes. When the virtual machine execution apparatus 1000 executes some processing, one or more classes are always loaded into the virtual machine engine 1400. If another class is used in the class function, the class being used is loaded.

  As described above, in the process at the time of startup, after initialization, the class file group is loaded from the secondary storage device 1200 and the link process is executed, and immediately after the virtual machine is started immediately before the class file is loaded. The state of the main memory 1300 is shown in FIG. At this point, there are only the virtual machine engine 1400 having an array for data management, and an operand 1420 including a primitive function processing routine and a primitive data type and its value. On the other hand, FIG. 5 shows the state of the main storage device 1300 after loading the class file group from the secondary storage device 1200 and executing the link processing. At this time, the processing procedure of the method indicated by the operation code (integer) in the file is replaced with a pointer to the processing function corresponding to the operation code. That is, processing can be realized by preparing an address management array (an address of an operation code in a data area in a class) as a method and sequentially executing function pointers stored in the array. In fact, there are multiple class files that call their own methods and methods other than self. Such a method call function can be called directly by the caller knowing the pointer to the function array managed by the method. Searching for an address on the real memory (main storage device 1300) as described above and recording the information on the use side is called link processing. These are performed immediately after startup, and only jump directly to the address when the actual processing starts. Therefore, this embodiment does not generate extra address calculation such as relative address calculation in the prior art. It is characterized by that.

  Further, as shown in FIG. 5, a method 1530 corresponding to each method of each method member 1510 includes a plurality of operation codes, addresses of which are described in the method, and each operation code execution module of the processing routine group 1410 Linked. FIG. 9 is a block diagram showing a class configuration in the virtual machine engine 1400 according to the present embodiment. In FIG. 9, when the class B opcode Q is quoted in the class A 1500A, it is linked by storing the address of the op code Q in the data area in the class A 1500A. Further, as shown in FIG. 5, each operand 1420 of the field member 1520 has a data type and a value format.

  FIG. 3 is a flowchart showing virtual machine execution processing executed by the central processing unit 1100 of FIG. FIG. 6 is a block diagram showing the configuration of the virtual machine execution apparatus 1000 during the execution process of the virtual machine of FIG.

  In FIG. 3, first, in step S11, a predetermined activation member function address is acquired from the class array of the virtual machine engine 1400, and in step S12, a corresponding runtime is generated. In step S13, the method frame 1610 is generated in the runtime 1600, the address of the member function to be executed is registered, and the program counter value of the method frame 1610 is initialized to 0 in step S14. In step S22, an initialization routine registered in the program counter value of the current method frame 1610 is acquired. In step S23, it is determined whether the processing routine 1410 is a function call. If YES, the process returns to step S13. If NO, the process proceeds to step S15. In step S15, it is determined whether or not the processing routine 1410 registered in the program counter value 1610 of the current method frame is a jump instruction. If YES, the process proceeds to step S16. If NO, step S17 is performed. Proceed to In step S16, the program counter value is increased or decreased according to the jump instruction, and then the process proceeds to step S19. On the other hand, in step S17, the processing routine 1410 is executed, the program counter value is incremented by 1 in step S18, and the process proceeds to step S19. In step S19, it is determined whether or not the member function is finished. If YES, the process proceeds to step S20. If NO, the process returns to step S22. In step S20, the method frame 1610 is released, and the process returns to the calling method frame. In step S21, it is determined whether or not the program is finished. If NO, the process returns to step S22. Terminates the execution process of the virtual machine.

  FIG. 6 illustrates a state of the main storage device 1300 when the virtual machine executes the class method. With reference to FIG. 6, the execution process of FIG. 3 will be described in detail below.

  In the runtime processing of the virtual machine, as described above, the processing routine group 1410 is executed from the class information generated by the runtime 1600 and the method frame 1610 and loaded into the main storage device 1400. First, the address of the activation member function of the activation class to be executed first is acquired, and the runtime 1600 is generated. One runtime 1600 is generated for one process flow. That is, when it is desired to execute by multitasking, a plurality of runtimes 1600 are generated. Here, a single task case will be described. Next, a method frame 1610 is generated, and the activation member function of the activation class acquired previously is registered. One method frame 1610 is prepared when one member function is executed. The method frame 1610 owns a program counter, a local variable array, and a stack area. As described above, the member function is an array in which the numbers assigned by the processing routine group 1610 are stored in the order of execution. However, the program counter indicates the extent of execution in this array. Local variable arrays are arrays of variables that are used only within member functions. In addition, the stack area is an area for storing a calculation result in the middle of processing routines being executed in sequence. Furthermore, when the method frame 1610 is generated and the activation member function is registered, the program counter is initialized. Processing routines 1410 registered in the function are executed in order. When the end of the function is reached, the execution of this method frame is complete. However, the processing routine 1410 has a function call function. When this function is executed, the member function of the corresponding class is acquired, a new method frame 1610 is generated and the address of the member function is registered, and then executed. Moves to the generated method frame. Similarly, the method frame generated by the function call advances the program counter while executing the processing routine, and at the same time as the function ends, the method frame is released and returned to the calling method frame. As described above, when all the function executions are completed and there are no remaining method frames, the program execution is completed.

  As described above, the runtime array at the time of execution of the virtual machine is generated at the stage of actually executing the method of the class, and information stored in the runtime 1600 is an array of an address to the activation method and a method frame. The reason why the virtual machine engine 1400 includes a runtime array for storing a plurality of runtimes 1600 is to realize multitask processing. The method frame 1610 generated in the runtime 1600 is generated every time a new method is executed. For example, when another method is activated in the activation method, two method frames are stored in the method frame array. The method frame 1610 stores an address to the execution method, a program counter indicating what number of operation code is being executed in the method, an address to the execution method, and at least one variable generated and used only in the method. The local variable area 1620 to be stored and the stack area 1630 that includes at least one operand and temporarily stores a calculation result in the middle of processing are stored as information. In the operation principle of method execution by the method frame 1610, when a method is first activated, a program counter value is initialized and a local variable is prepared. Then, the execution method address is accessed, and an array in which the opcode addresses are stored in the execution order is executed according to the program counter value. The progress of operations on local variables is stored in the stack area. Each time the processing of the operation code is completed, the program counter value is incremented by one and the next operation code is executed. In addition, the operation code representing the jump instruction is realized by increasing or decreasing the program counter, and unnecessary address calculation is unnecessary.

  In this embodiment, the field member 1520 generated for each class 1500, and the operand 1420 storing the data tag and data value in the local variable area 1620 and stack area 1630 generated for each method frame 1610. Is formed in a format including a primitive data type and its value. Specifically, the local variable area 1620, the stack area 1630, and the field member 1520 of each class included in the method frame 1610 are characterized by an array of operands. Normally, the contents to be recorded in the stack area store only data values. For this reason, in the primitive instruction of the virtual machine, instructions are prepared by distinguishing data types such as iadd which is addition for integers and fadd which is addition for floating point, and the stack values are manipulated. Therefore, the operation code is increased, but in the virtual machine according to the present embodiment, as shown in FIG. 8, the data code for distinguishing the data type is also arranged in the stack area 1630 and the code is unified. The instruction system has been simplified.

  FIG. 7 is a table showing an example of an opcode used in the virtual machine execution apparatus 1000 according to the present embodiment. As shown in FIG. 7, the operation code is represented by a non-negative integer. For example, if 0x0001, an addition process is executed on the virtual machine, and if 0x0002, a subtraction process is executed on the virtual machine. The actual processing contents exist as a processing routine group 1410 for each operation code, such as an execution module 1410a for the operation code a, an execution module 1410b for the operation code b, an execution module 1410c for the operation code c, and the like. Exists. At the stage where the class intermediate code 1201 is recorded in the secondary storage device, the class method records the ID of the operation code, but changes the ID of the operation code to the address to the execution module at the timing of loading to the memory. For example, if the first execution of the method is the operation code c, it is converted to the address of the execution module 1410c of the operation code. This eliminates the search by ID during execution and enables high-speed execution. Further, the address calculation such as (reference address + offset) is made unnecessary by directly describing the address of the main storage device 1300.

  FIG. 8 is a table showing an example of data codes used in the virtual machine execution apparatus 1000 according to the present embodiment. As shown in FIG. 8, the data code is represented by a non-negative integer. For example, 0x0002 indicates a 2-byte integer, and 0x0003 indicates a 4-byte integer. In the virtual machine according to the present embodiment, the instruction system is simplified by arranging the data code for distinguishing the data type in the stack area 1630 and unifying the codes.

Modified example.
FIG. 10 is a block diagram showing a configuration of a virtual machine execution device 1000A according to a modification of the present invention. In the virtual machine execution device 1000 according to the embodiment of FIG. 1, the class intermediate code stored in the secondary storage device 1200 provided in the virtual machine execution device 1000 is loaded, expanded into the main storage device 1300, and executed. However, the present invention is not limited to this, and may be configured as the virtual machine execution device 1000A in FIG. That is, the virtual machine execution apparatus 1000A includes a communication interface circuit 1710. For example, the class intermediate code is transmitted from the secondary storage device in the server apparatus 1900 of the digital computer remotely via the network 1800 such as the Internet via the network 1800. It may be executed after being downloaded to the main storage device 1300 and expanded.

  In the above embodiment, the runtime 1600 is configured to be able to reserve a local variable area 1620 that stores an address to an execution method and a variable that is generated and used only within the method, and the runtime 1600 May include a stack area 1630 for temporarily storing calculation results during processing, and the stack area 1630 may be configured to store an operand 1420 expressed in the form of a data type and its value.

  As described above in detail, according to the present invention, when the intermediate code is loaded into the main storage device, the address of the operation code included in the method is represented by the address of the main storage device, and the data area of each class. Therefore, it is not necessary to perform an address conversion process when the virtual machine is executed, so that the virtual machine can be executed at a higher speed than in the prior art.

  Further, according to the present invention, since the operand stored in the stack area is expressed in the form of its data type and its value, a virtual machine having a compact specification can be suppressed without increasing the instruction code due to the combination of the data types. Can be realized.

It is a block diagram which shows the structure of the virtual machine execution apparatus 1000 which concerns on one Embodiment of this invention. 2 is a flowchart showing virtual machine installation processing executed by the central processing unit 1100 of FIG. 1. 3 is a flowchart showing virtual machine execution processing executed by the central processing unit 1100 of FIG. 1. FIG. 3 is a block diagram showing a configuration of a virtual machine execution device 1000 after the process of step S1 of FIG. FIG. 3 is a block diagram showing a configuration of a virtual machine execution device 1000 after the process of step S3 of FIG. FIG. 4 is a block diagram illustrating a configuration of a virtual machine execution device 1000 during the virtual machine execution process of FIG. 3. It is a table | surface which shows an example of the opcode used with the virtual machine execution apparatus 1000 which concerns on this embodiment. It is a table | surface which shows an example of the data code used with the virtual machine execution apparatus 1000 which concerns on this embodiment. It is a block diagram which shows the structure of the class in the virtual machine engine 1400 which concerns on this embodiment. It is a block diagram which shows the structure of 1000A of virtual machine execution apparatuses which concern on the modification of this invention.

Explanation of symbols

1000 ... virtual machine execution device,
1100: Central processing unit,
1200 ... secondary storage device,
1201 ... Class intermediate code,
1300 ... main memory,
1400: Virtual machine engine,
1410 ... Processing routine group,
1410a ... Opcode a execution module,
1410b ... Opcode b execution module,
1410c ... Opcode c execution module,
1420 ... Operand,
1500, 1500A, 1500B ... class,
1510 ... Method member,
1520 ... Field members,
1530 ... Method,
1600 ... runtime,
1610 ... Method frame,
1620 ... Local variable area,
1630 ... Stack area,
1700 ... Bus
1710: Communication interface circuit,
1800 ... network,
1900: Server device.

Claims (6)

In a virtual machine execution device that loads and executes an intermediate code from a secondary storage device to a main storage device,
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
A virtual machine execution device characterized in that when the intermediate code is loaded into a main storage device, an address of an operation code included in the method is represented by the address of the main storage device and stored in the data area of each class. The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
2. The virtual machine execution device according to claim 1, wherein operands expressed in a data type and a value format are stored in the stack area. In a virtual machine execution device that loads and executes an intermediate code from a secondary storage device to a main storage device,
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
A virtual machine execution device, wherein an operand expressed in a data type and a value format is stored in the stack area. In a virtual machine execution method for loading and executing an intermediate code from a secondary storage device to a main storage device,
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
A virtual machine including a step of storing an address of an operation code included in the method by an address of the main storage device and storing it in the data area of each class when the intermediate code is loaded into the main storage device Execution method. The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
5. The virtual machine execution method according to claim 4, wherein operands expressed in a data type and a value format are stored in the stack area. In a virtual machine execution method for loading and executing an intermediate code from a secondary storage device to a main storage device,
A virtual machine engine comprising a class array having at least one class loaded with intermediate code in main memory and a runtime array having at least one runtime for storing runtime information;
The class includes a method member having at least one method describing processing procedure information including an operation code, and a field member having at least one reference information to numeric data or character string data or another class,
The runtime can secure an address to the execution method and a local variable area for storing variables that are generated and used only within the method, and has a stack area for temporarily storing calculation results during processing. ,
A virtual machine execution method characterized in that an operand represented by a data type and a value format is stored in the stack area.

Images